New method developed to identify major scientific breakthroughs
開發出識別重大科學突破的新方法
Scientists from Binghamton University and the University of Virginia have introduced a groundbreaking machine-learning method to identify major scientific breakthroughs.
來自賓漢頓大學（ㄅㄧㄣㄏㄢˋㄉㄨㄣˋㄉㄚˋㄒㄩㄝˊ）與維吉尼亞大學（ㄨㄟˊㄐㄧˊㄋㄧˊㄧㄚˋㄉㄚˋㄒㄩㄝˊ）的科學家們，推出了一種突破性的機器學習方法，用以識別重大的科學突破。
Published in Science Advances, this approach moves beyond traditional citation counts to measure how research redirects a field's trajectory.
這項發表在《科學前沿》（ㄎㄜㄒㄩㄝˊㄑㄧㄢˊㄧㄢˊ）期刊上的研究，超越了傳統引用計數，旨在衡量研究如何改變該領域的軌跡。
The team utilizes neural embedding to map millions of papers and patents using two vectors: intellectual lineage and the cascade of influence.
該團隊利用神經嵌入（ㄕㄣˊㄐㄧㄥㄧㄣˋㄖㄨˋ）技術，透過智力傳承（ㄓˋㄌㄧˋㄔㄨㄢˊㄔㄥˊ）與影響瀑布（ㄧㄥˇㄒㄧㄤˇㄆㄨˋㄅㄨˋ）這兩個向量，對數以百萬計的論文和專利進行繪圖。
By analyzing vast datasets, the method provides a scalable, dynamic map of innovation.
透過分析龐大的數據集，此方法提供了一張具備可擴展性且動態的創新地圖。
Beyond academic interest, this tool has major practical implications for science policy.
除了學術興趣外，此工具對科學政策具有重大的實務影響。
It helps funding agencies identify emerging, high-impact fields and allows historians to map the social dynamics behind paradigm shifts.
它有助於資助機構識別新興且具高影響力的領域，並讓歷史學家能夠繪製出範式轉移（ㄈㄢˋㄕˋㄓㄨㄢˇㄧˊ）背後的社會動態。
By successfully identifying historical Nobel Prize-winning studies, the researchers have validated the model's ability to spot true innovation.
研究人員透過成功識別歷史上獲得諾貝爾獎的研究，證實了該模型識別真正創新能力的有效性。
